Manage customer enquiries from Telegram
How I built an AI customer support system with n8n that handles emails, tasks, and replies automatically from Telegram. Task: The author built an n8n workflow to centralize customer enquiries from emails and forms. The system captures enquiries, assigns unique IDs, creates ClickUp tasks, and sends Telegram notifications. The human replies via Telegram, and AI determines the next action (send email, update status, or close thread) using OpenAI, Gmail, and Google Sheets.; tools: n8n. Human role: The original post does not say how a person stayed involved.; limitations: Prerequisites include accounts for n8n, OpenAI, Telegram, ClickUp, Gmail, and Google Sheets. The source does not specify costs, API limits, or error handling for failed AI decisions. Human approval is implicit in the Telegram reply step, but the mechanism for overriding AI actions is not detailed. Source reliability is secondary, based on a social media post..
Tools
Published on Agent Work Atlas:
Overview
How I built an AI customer support system with n8n that handles emails, tasks, and replies automatically from Telegram.
Workflow
The author built an n8n workflow to centralize customer enquiries from emails and forms. The system captures enquiries, assigns unique IDs, creates ClickUp tasks, and sends Telegram notifications. The human replies via Telegram, and AI determines the next action (send email, update status, or close thread) using OpenAI, Gmail, and Google Sheets.
Reported result
The source does not report measured outcomes, efficiency gains, or error rates. It claims the workflow prevents context loss and reduces tool switching, but these are qualitative assertions without data. A measurable acceptance check would be tracking the reduction in average response time or the percentage of enquiries resolved without manual context retrieval.
Human role
The original post does not say how a person stayed involved.
Limitations
Prerequisites include accounts for n8n, OpenAI, Telegram, ClickUp, Gmail, and Google Sheets. The source does not specify costs, API limits, or error handling for failed AI decisions. Human approval is implicit in the Telegram reply step, but the mechanism for overriding AI actions is not detailed. Source reliability is secondary, based on a social media post.
Try it
I want to build an AI-assisted customer support workflow using n8n. My inputs are [replaceable: email address for enquiries, Telegram bot token, ClickUp project ID, OpenAI API key]. The goal is to capture enquiries, create tasks, and allow me to reply via Telegram while AI handles the final email/status update. Steps: 1. Set up an n8n workflow triggered by new emails in [replaceable: Gmail inbox]. 2. Use OpenAI to extract key info and assign a unique ID. 3. Create a ClickUp task with this ID and send a Telegram notification to me. 4. When I reply on Telegram, use OpenAI to classify the intent (send email, update status, close). 5. Execute the action via Gmail or ClickUp APIs. Deliverables: A JSON export of the n8n workflow and a screenshot of a successful test run. Quality check: Ensure the unique ID persists across all tools. Safeguards: Run a dry test with one sample enquiry first. Do not send real emails or update live tasks until I confirm the AI's decision logic is correct. If inputs are missing, ask me for the specific API keys or IDs needed.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.